全 文 ：J. Geogr. Sci. (2008) 18: 201-210
DOI: 10.1007/s11442-008-0201-7
© 2008 Science in China Press Springer-Verlag

Received: 2007-12-17 Accepted: 2008-02-28
Foundation: National Natural Science Foundation of China, No.30270227; No.90102005; Knowledge Innovation Project
of the CAS, No.KZCX3-SW-321; No.KZCX1-10-02
Author: Wu Pu (1979−), Ph.D Candidate, specialized in global change. E-mail: wup@igsnrr.ac.cn
www.scichina.com www.springerlink.com

Reconstruction of summer temperature variation
from maximum density of alpine pine during
1917−2002 for west Sichuan Plateau, China
WU Pu, WANG Lily, SHAO Xuemei
Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China

Abstract: Having analyzed the tree ring width and maximum latewood density of Pinus den-
sata from west Sichuan, we obtained different climate information from tree-ring width and
maximum latewood density chronology. The growth of tree ring width was responded princi-
pally to the precipitation in current May, which might be influenced by the activity of southwest
monsoon, whereas the maximum latewood density reflected summer temperature
(June–September). According to the correlation relationship, a transfer function had been
used to reconstruct summer temperature for the study area. The explained variance of re-
construction is 51% (F=52.099, p<0.0001). In the reconstruction series: before the 1930s, the
climate was relatively cold, and relatively warm from 1930 to 1960, this trend was in accor-
dance with the cold-warm period of the last 100 years, west Sichuan. Compared with
Chengdu, the warming break point in west Sichuan is 3 years ahead of time, indicating that
the Tibetan Plateau was more sensitive to temperature change. There was an evident sum-
mer warming signal after 1983. Although the last 100-year running average of summer tem-
perature in the 1990s was the maximum, the running average of the early 1990s was below
the average line and it was cold summer, but summer drought occurred in the late 1990s.
Keywords: west Sichuan (province) plateau; tree ring; summer temperature
1 Introduction
West Sichuan Plateau, a major part of the southeast Qinghai–Tibet Plateau, is situated be-
tween the Qinghai–Tibet Plateau and Sichuan Basin, and takes up 2/3 of the area of Sichuan
Province. Researches indicated that, from the 1980s to 1990s, the temperature in Sichuan
Basin dropped while global warming is an acknowledged fact. Therefore, research on cli-
mate change of Sichuan Province has aroused the interest of the researches, who have fo-
cused their efforts on this special phenomenon. So scholars hold generally that studies on the
characteristics of temperature and precipitation variations in west Sichuan is of great sig-
nificance to the understanding of the regional climate change of the Qinghai–Tibet Plateau
202 Journal of Geographical Sciences

and Sichuan Province under the background of global warming and even the differences of
regional response to global warming (Li, 2003).
Increasing concerns on future climate changes and their potential on humankind have
prompted scientists to explore a variety of historical and natural archives that have reference
value for past climate changes. Among these tree-ring data play a prominent role in the char-
acterization and assessment of climate variations prior to the instrumental period, for its
strength points: continuity and precise datability, and easily replication (Fritts, 1976). As a
result, tree ring is the preferred proxy for past climate change research and is an important
technical approach on PAGES (Past Global Changes) (Fritts, 1976; LaMarche, 1978).
The research on dendroclimatology in China started in the 1930s, focusing on mainly tree
ring width with less efforts on tree ring density. The dendroclimatology in China has made
great progress since the 1990s. However, the emphases were mainly placed on the Qing-
hai–Tibet Plateau and typical arid/semiarid areas in Northwest China (Wu et al., 1988;
Zhang et al., 2003; Gou, 2004; Shao et al., 2003, 2004), whereas it was inadequately or ab-
sent in tropical and sub-tropical zones due to the growth conditions for trees are not
obviously limited (Sun and Zhong, 1997, 1999; Shao and Fan, 1999; Qian et al., 2001; Chen
et al., 2002; Xing et al., 2004). International dendroclimatic studies are mainly reported in
Indonesia and South America where the dry-wet seasonal changes are obvious (D’Arrigo et
al., 1994; Villalba et al., 1998; Worbes, 1999).
Wood density, reflecting indirectly cell structure of tree rings such as cell diameters and
cell-wall thickness of xylem cell and cell cavity structure, may contain abundant information
in characteristics of tree-ring growth compared those in tree ring width. In particular, on up-
per treeline and high-latitude regions, tree-ring maximum density is rather sensitive to tem-
perature changes, and can be used for reconstructing temperature (Jacoby and D’Arrigo,
1989; Wang et al., 2000; Esper, 2000; Briffa et al., 2002; Esper et al., 2002a, 2002b, 2003).
Alpine pine (Pinus densata Mast) is a photophobic tree species with tolerance of cold,
arid and leanness. It plays an important role in lower treeline of sub-alpine coniferous forest
in western Sichuan; furthermore, it is able to form predominant community at section of lean
sunny slope. Also, it is one of the zonal vegetation types in the middle of the Hengduan
Mountains (Collaboration Group of Sichuan Vegetation, 1980). The formation and distribu-
tion of stable vegetation community of alpine pine indeed implicate climatic information.
This paper studies the relationship between tree ring width and density and climatic fac-
tors by the samples collected from west Sichuan Plateau, which is the modern distribution
center of subalpine ever-green coniferous forest in China (Collaboration Group of Sichuan
Vegetation, 1980). In this paper, we reconstructed summer temperature (June–September)
from maximum latewood density of alpine pine in western Sichuan. It contributes to under-
stand better the climate change of the Qinghai–Tibet Plateau and Sichuan Province, even the
abnormity of regional response to global warming.
2 Materials and methods
2.1 Study area
Samples were collected at middle tree line for 49 cores out of 28 trees using an increment
borer at Sesiman (sm) (31.32°N, 101.99°E, 2805 m asl) and 35 cores from 20 trees at upper
tree line of alpine pine Xiandiaogou (xd) (31.5°N, 102°E, 3234 m asl) in Jinchuan County
(Figure 1).
WU Pu et al.: Reconstruction of summer temperature variation from maximum density of alpine pine 203





Figure 1 The location of sample sites and meteorological station

The sample sites are located at the margin of southeast Qinghai–Tibet Plateau. The to-
pography of this region is characterized by extremely steep high-mountain physiognomy and
intensively dissected by deep gorges of rivers with a height of from 1500 m up to 5300 m
and a relative height difference of 2000–2500 m. Topographically, it descends from north-
west to southeast.
The climate in the study area is mainly influenced by the south branch of the westerlies,
Southeast and Southwest Asian monsoon. The mean annual temperature is below 10℃; an-
nual precipitation over 700 mm; and the mean frost-free period at Jinchuan is about 170
days. The weather there is characterized by cool and rainy summer and cold and sunny win-
ter.
In terms of the vegetation regionalization, it belongs to the middle and upper Dadu River
vegetation plots of the west Sichuan alpine-valley coniferous forest zone in the subtropical
evergreen broadleaved forest region.
2.2 Climate data
The instrumental records were collected from the meteorological station in Xiaojin (31°N,
102°E, 2369.2 m asl), the nearest station to the sample sites. Monthly mean temperature and
precipitation were used from previous October to the current September. The climate data
after September were not used because alpine pine has finished its radial growth at that time
and comes to period of dormancy (Zheng, 1983). From Figure 2, annual temperature appears
mild in range and smooth in variation, characterized by cool summer and not-so-cold winter.
The maximum temperature reaches an average of 20–22℃ in July and a minimum of 1.5℃
in January. Annual precipitation change curve shows double peak type, June and September,
respectively. The rainy season is from May to September, and dry season from November to
204 Journal of Geographical Sciences




Figure 2 Monthly mean temperature and total precipitation at Xiaojin meteorological station

next March. The precipitation in August decreases in rainy season may be caused by the
northward moving troughs (Ye, 1952).
2.3 Cross-dating and chronology development
After the increment cores were air-dried, glued to wooden mounts, and sanded in the labo-
ratory, the skeleton-plot technique was utilized for the primary crossdating, and the ring
width was subsequently measured to the nearest 0.01 mm with dendrometer. The accuracies
of crossdating and measurements were then checked using the COFECHA computer pro-
gram (Holmes, 1983). We used 32-year spline function to filter low-frequency signal in each
tree-ring width series, and then verified the detrended series with the lag of half of the mean
length of series because the length of all series were no less than one hundred years. Ac-
cordingly, the length of running rbar was adopted as the half of the lag (Grission-mayer,
2001).
After ring width series were crossdated, we cut the cores into small pieces along with the
xylem angular and glued to wooden mounts, then cut to 1 mm lath with DENDROCUT.
X-ray film was taken for those laths. We measured the film with DENDRO2003 to obtain 7
parameters including tree ring width (TRW), earlywood width (EWW), latewood width
(LWW), earlywood mean density (EWD), latewood mean density (LWD), minimum density
(MID) and maximum density (MXD) at tree ring laboratory in Laval University, Canada.
The chronologies were developed by ARSTAN program. Observing the curve of the raw
width and density series, we didn’t find that the width curve of alpine pine shows obvious
age trend. Similarly, the density curve fluctuates around the average without growth trend
related to age. We used 30-year spline function and horizontal line to fit the age trend. The
detrended series were then averaged into site chronology using the method of bi-weight ro-
bust estimation of mean (Cook and Kairiukstis, 1989). Because detrended series were ob-
tained by taking the ratio of the measurement over the fitted value in each year, the standard
ring width and density chronologies are dimensionless series with a mean of 1 and minimum
of 0. According to the experience from developing chronology in arid and semiarid region,
WU Pu et al.: Reconstruction of summer temperature variation from maximum density of alpine pine 205


besides chronology, we compared standard, residual and arstan chronologies and finally cho-
se the residual chronology as the indicator of radial growth (Table 1).

Table 1 Chronological statistics for GSxd and GSsm
Site Chronology Mean Mean sensitivity Standard deviation Auto correlation
STD 1.017 0.062 0.091 0.347
RES 1.005 0.080 0.073 −0.061 GSxd
ARS 1.005 0.062 0.078 0.345
STD 0.989 0.051 0.055 0.306
RES 0.998 0.053 0.046 −0.146 GSsm
ARS 0.997 0.047 0.050 0.254

3 Response of radial growth to climate
We used response function and correlation analysis to identify possible relationships be-
tween monthly mean temperature and precipitation from previous October to the current
September and TRW and MXD series. The results of response function analyses indicate
that the precipitation of current May strongly influenced tree-ring width, whereas the maxi-
mum density was closely correlated with the temperature of previous October, the current
June to September and the precipitation of the current July to September as well. The corre-
lation analyses of MXD chronology and climate (Figure 3) suggest that MXD was positively
correlated with the temperature of June to September during the current growing season, and
negatively correlated with the precipitation of previous October and the current July to Sep-
tember. Besides, the MXD of GSsm (alpine pine in the site of sm) and GSxd (alpine pine in
the site of xd) is significant with the temperature of June to September and the precipitation
of June to August in growing season at 99% confidence level, moreover, the MXD of GSsm



Figure 3 The response function analyses of MXD residual chronologies
Note: * denotes the significance at 95% level; ** denotes the significance at 99% level
206 Journal of Geographical Sciences

is significant with the temperature of the previous October at 95% confidence level. Favorite
temperature contributes to photosynthesis, cell split and cell-wall thickening from June to
September in growing season when tree grows vigorously, thus generating higher MXD.
Partial correlation analysis shows that there is no significant relationship between MXD and
precipitation when the temperature variable is controlled. The negative correlation between
MXD and precipitation is due to the existence of significant negative correlation between
MXD and temperature at this area. The physiological mechanism of negative correlation
between MXD and previous October is unknown and needs further study.
4 Transfer function
According to the results of correlation and partial correlation analyses, we chose monthly
mean temperature as variable for reconstruction; furthermore, we combined factors which
have the same influences on tree growth, and then conducted regression analysis between
MXD and the combined variables. The result shows that summer temperature (June to Sep-
tember) is a most sensitive factor to MXD chronology, which is consistent with the conclu-
sion of D’Arrigo et al. (1996). Therefore, we reconstructed summer temperature (June to
September) using maximum latewood density chronology. The correlation coefficient be-
tween Sesiman and Xiandiaogou chronology is so low (r=0.508) at 99% confidence level
that we can’t use the mean series of the two chronologies as regional chronology. Instead,
we analyzed the two chronologies by principal component (PC). The PC analyses show that
the first PC explained 75.4% of the total and the second PC the rest 24.6%. As a result, we
eventually chose the first PC as variable for reconstructing summer temperature (June to
September). A transfer function is designed as
T=1.834PCA1+73.409
where T presents summer temperature (June to September) and PCA1 is the first PC.
As shown in Table 2, the transfer function explained 52% of the variance of summer
temperature. Neglecting the influence of freedom, it still accounts for 51%, the F value

Table 2 Transfer function and cross-verification statistics
Vari-
ance/%
Adjusted
variance/% F-value
First-order
signal test
Signal
test
Product
means
Reduced
error
Correlation
coefficient
52 51 52.099 34** 34** 3.9304 0.4629 0.6814



Figure 4 Comparison between the reconstructed summer (June–September) temperature and the actual instru-
mental data in west Sichuan
WU Pu et al.: Reconstruction of summer temperature variation from maximum density of alpine pine 207


52.099 exceeds the 99.99% confidence level. Comparing the instrumental temperature and
the reconstructed temperature by the first PC, we found high agreement in low frequency as
well as in high frequency, and there are no singular points. All those suggested the equation
was stable (Figure 4).
5 Summer temperature reconstruction
Based on the transfer function, the summer temperature (June to September) was recon-
structed for the period of 1917 to 2002 as shown in Figure 5. The research period was lim-
ited to AD 1917−2002 in this study to ensure the reliability of climate reconstruction with
enough samples depth. After 15-year moving mean (the thick line in Figure 5), the correla-
tion coefficient between reconstructed and instrumental climate is the highest (r=0.766). All
statistics suggests the low-frequency of reconstruction is reliable.



Figure 5 The temperature reconstructed based on the MXD from June to September during 1917 to 2002 for the
west Sichuan Plateau (The thick line is the 15-year moving average and the straight line is annual mean.)

It is obvious that several interannual and interdecadal fluctuations occurred during the
past nearly 100 years. From the reconstruction, we found that summer temperature in this
area tended to increase during the early 20th century, and it reached a peak during the
mid-1920s, slightly ahead of the 1930s’ peak of summer temperature in China (Lin et al.,
1995). In general, the climate was relatively cold before the 1930s, and relatively warm from
1930 to 1960, this tendency was accordance with the cold–warm period of the last 100 years
in west Sichuan (Chen et al., 1999). It is meaningful to compare the reconstructed summer
mean temperature in Xiaojin with the annual mean temperature in Sichuan for the significant
correlation (r=0.571) at 99% confidence level between them. Then it began to decrease and
increased again since 1983. Although the year 1983 is not the abrupt point according to veri-
fication by Yamamoto’s method (Fu and Wang, 1992; Yu, 2004) (Table 3), when compared
with Chengdu, the warming break point in west Sichuan is 3 years ahead of time (Chen et al.,
1999). The result showed Tibetan Plateau was more sensitive to temperature change and
west Sichuan Plateau may be the presymptom area of climate change in Sichuan (Li and He,
1999). There was an evident summer warming signal after 1983. Although the last 100-year
moving average of summer temperature in the 1990s was the maximum, yet that of the early
1990s is still below the average line and it was cold summer; but summer drought occurred
in the late 1990s (Yang et al., 1994).
208 Journal of Geographical Sciences

Table 3 Detection of abrupt climate change in 1983
a=10 a=15 a=16
S/N 0.094 0.034 0.07
True or false False False False
6 Discussion
Generally, trees growing in the upper forest line are primarily controlled by temperature
(Fritts, 1976). In this study area, although the climate is complacency for tree growth, the
MXD of alpine pine is controlled by summer temperature. Compared to tree-ring width, tree
ring density reflects not only the influence of climate change annually, but also in-
ter-annually. We can extract more climate information from the residual MXD chronology
than that from tree ring width chronology.
Although precipitation is usually considered as insignificant to tree growth (LaMarche,
1974), the precipitation of May significantly correlates to tree ring width of alpine pine in
this study which can be explained biologically. The southwest monsoon in the summer is the
main source of precipitation in this study area, which moves northwards along with the
southeastern slope of the Qinghai–Tibet Plateau during May and June, then rotates cycloni-
cally at the eastern margin, and the convergence line between the sir current from north and
the monsoon thus brings rain. Therefore, the radial growth is quite sensitive to the arriving
time of rain in the period of growing season of alpine pine.
Since the height growth of alpine pine ends in June (Zheng, 1983), the photosynthesis of
total leaves is most sufficient for current July to September. At high latitude region of the
northern hemisphere, the thickening of latewood cell-wall wouldn’t influence the space of
cell protoplasm (Schweingruber and Fritts, 1988; Anatova and Stasova, 1993; Vaganov,
1996), so that we can consider the maximum latewood density as net photosynthesis. Longer
growing season and higher accumulative temperature offer more time for the thickening of
cell-wall, resulting in larger maximum latewood density. And it shows statistically that
maximum latewood density significantly positive-correlated to the temperature summation
from June to September. Precipitation in July and August might indirectly affect maximum
latewood density. The variations of precipitation and relative humidity in July and August in
Xiaojin are accordant and they are strongly negative-correlated to temperature. More pre-
cipitation means higher relative humidity and fewer sunlight days, decreasing indirectly the
production accumulation of net photosynthesis, thus lowering the maximum latewood den-
sity.
7 Conclusions
The following conclusions can be drawn.
(1) The climate information extracted from tree ring width and density chronologies is
different for alpine pine from two sites (Sesiman and Xiandiaogou) in Xiaojin; tree-ring
density variables in this region seemed effective to evaluate climate variations, and showed
great potential in reconstructing climate change.
(2) The precipitation in current May strongly influences tree ring width, whereas the
WU Pu et al.: Reconstruction of summer temperature variation from maximum density of alpine pine 209


maximum density indicates the temperature fluctuation during current June to September.
(3) The summer temperature is reconstructed (June to September) for the past 100 years
in Xiaojin with transfer function. The reconstruction can explain 51% of the total variances
(F=52.099, p<0.0001). According to the reconstruction, summer temperature in this area
tended to increase during the early 20th century, and it reached a peak during the mid-1920s,
slightly ahead of the 1930s’ peak of summer temperature in China. In general, the climate
was relatively cold before the 1930s, and relatively warm from 1930 to 1960, and then it
began to decrease, and then increased again since 1983. Compared to Chengdu, the warming
break point in west Sichuan is 3 years ahead of time, indicating that the Tibetan Plateau was
more sensitive to temperature change and west Sichuan may be the presymptom area of cli-
mate change in Sichuan Province. There was an evident summer warming signal after 1983.
Although the last 100-year running average of summer temperature in the 1990s was the
maximum, the running average of the early 1990s was below the average line and it was
cold summer; summer drought occurred in the late 1990s.
References
Anatova F A, Stasova V V, 1993. Effects of environmental factors on wood formation in Scots pine stems. Trees,
7(4): 214–219.
Briffa K R, Osborn T J, Schweingruber F H et al., 2002a. Tree-ring width and density data around the Northern
Hemisphere (Part 1): Local and regional climate signals. The Holocene, 12(6): 737–757.
Chen Baojun, Qian Junlong, Pu Peimin, 2002. Relationship between ENSO and tree-ring table isotopes of Cryp-
tomeria fortunei from Tianmu Mountain, East China. Marine Geology & Quaternary Geology, 22(4): 53–58.
(in Chinese)
Chen Wenxiu, Hao Kejun, Tian Hong, 1999. Multiple time scale analyses of surface temperature variations in
Sichuan during last 100 years. Journal of Chengdu Institute of Meteorology, 14 (1): 78–82. (in Chinese)
Collaboration Group of Sichuan Vegetation, 1980. Sichuan Vegetation. Chengdu: Sichuan People’s Publishing
House. (in Chinese)
Cook E R, Kairiukstis L A, 1989. Methods of Dendrochronology. Kluwer Academic Publishers, 125.
D’Arrigo R D, Cook E R, Jacoby G C, 1996. Annual to decadal-scale variations in northwest Atlantic sector tem-
perature inferred from Labrador tree-ring. Canadian Journal of Forest Research, 26: 143–148.
D’Arrigo R D, Jacoby G C, Krusic P J, 1994. Progress in dendroclimatic studies in Indonesia. TAO, 15(3):
349–363.
Esper J, 2000. Long term tree-ring variations in Junipers at the upper timberline in the Karakorum (Pakistan). The
Holocene, 10: 253–260.
Esper J, Cook E R, Schweingruber F H, 2002a. Low-frequency signals in long tree-ring chronologies for recon-
structing of past temperature variability. Science, 295: 2250–2253.
Esper J, Schweingruber F H, Winiger M, 2002b. 1300 years of climatic history for Western Central Asia inferred
from tree-rings. The Holocene, 12: 267–277.
Esper J, Shiyatov S G, Mazepa V S et al., 2003. Temperature-sensitive Tien Shan tree ring chronologies show
multi-centennial growth trends. Climate Dynamics, 8: 699–706.
Fritts H C, 1976. Tree Rings and Climate. New York: Academic Press, 228–236.
Fu Congbin, Wang Qiang, 1992. The conception and detection methods of abrupt climate change. Chinese Jour-
nal of Atmospheric Sciences, 6(4): 482–493. (in Chinese)
Gou Xiaohua, Chen Fahu, Yang Meixue et al., 2004. Analysis of the tree-ring width chronology of Qilian Moun-
tains at different elevation. Acta Ecologica Sinica, 24(1): 172–176. (in Chinese)
210 Journal of Geographical Sciences

Grission-mayer H D, 2001. Evaluating crossdating accuracy: A manual and tutorial for the computer program
cofecha. Tree-ring Research, 57(2): 205–221.
Jacoby G C, D’Arrigo R D, 1989. Reconstructed northern hemisphere annual temperature since 1671 based on
high-latitude tree ring data from North America. Climatic Change, 14: 39–59.
LaMarche Jr. V C, 1974. Frequency-dependent relationships between tree-ring series along an ecological gradient
and some dendrochronological implications. Tree-Ring Bulletin, 34: 1–20.
LaMarche, 1978. Tree-ring evidence of past climatic variability. Nature, 276(5686): 334–338.
Li Chuan, Chen Jing, Zhu Yanjun, 2003. Preliminary research on climate change of west Sichuan Plateau in recent
fifty years. Plateau Meteorology, 22(suppl.): 128–144. (in Chinese)
Li Guoping, He Wenbin, 1999. Analyses of climatic fluctuation in the past 50 years and forecast of change trend
in future in Sichuan. Journal of Chengdu Institute of Meteorology, 9(3): 72–78. (in Chinese)
Lin Xuechun, Yu Shuqiu, Tang Guoli, 1995. Series of average air temperature over China for the last 100-year
period. Scientia Atmospherica Sinica, 19(5): 525–534. (in Chinese)
Qian Junlong, Deng Ziwang, Tu Qipu, 2001. Climatic significance of δD time series in tree rings from Tianmu
Mountain. Science in China (Series D), 31(5): 372–376.
Schweingruber, Fritts H C, 1988. Tree Rings. Dreidel Publishing Company.
Shao Xuemei, Fan Jinmei, 1999. Past climate on west Sichuan Plateau as reconstructed from ring-widths of
dragon spruce. Quaternary Sciences, 1: 81–89. (in Chinese)
Shao Xuemei, Fang Xiuqi, Liu Hongbin et al., 2003. Dating the 1000-year-old Qilian Juniper in mountains along
the eastern margin of the Qaidam Basin. Acta Geographica Sinica, 58 (1): 90–100. (in Chinese)
Shao Xuemei, Huang Lei, Liu Hongbin et al., 2004. Reconstruction of precipitation variation from tree rings in
recent 1000 years in Delingha, Qinghai. Science in China (Series D), 34(2): 145–153.
Sun Fan, Zhong Zhangcheng, 1997. Study on the power spectrum of growth ring of Journal of Gordonia acumi-
nate in Mt. Jinyun. Journal of Southwest Agricultural University, 19(1): 25–28. (in Chinese)
Sun Fan, Zhong Zhangcheng, 1999. Relationship between tree ring growth of Gordonia acuminate and climatic
factors in Mt. Jinyun. Chinese Journal of Applied Ecology, 14(2): 151–154. (in Chinese)
Vaganov E A, 1996. Analysis of seasonal tree-ring formation and modeling in dendrochronology. In: Dean J D,
Meko D M, Swetnam T W (eds.). Tree Rings, Environment and Humanity, Radiocarbon, 73–87.
Villalba R, Grau H R, Boninsegna J A, 1998. Tree-ring evidence for long-term precipitation changes in subtropi-
cal South America. International Journal of Climatology, 18(3): 1463–1478.
Wang L, Payette S, Begin Y, 2000. A 1300-year tree-ring width and density series based on living, dead and sub-
fossil black spruce at tree-line in Subarctic Quebec, Canada. Holocene, 11: 333–341.
Worbes M, 1999. Annual growth rings, rainfall-dependent growth and long-term growth problems of tropical trees
from the Caparo Forest Reserve in Venezuela. Journal of Ecology, 87(3): 391–403.
Wu Xiangding, Sun Li, Chen Zhigang, 1988. The development of several tree ring chronology in Tibetan Plateau.
Chinese Science Bulletin, 8: 616–619. (in Chinese)
Xian Qiuru, Liu Hongyan, Sun Yanrong et al., 2004. Response of contemporary tree ring width of Cinnamomum
camphora to climate change in Yangchun, Guangdong Province. Acta Ecologica Sinica, 24(9): 2076–2080. (in
Chinese)
Yang Fu’an，Xu Yujiang, Chen Xiaomeng, 1994. Characteristic analysis of precipitation and drought or excessive
rain in Sichuan Province near 50-year and forecast for the coming 10 years. Sichuan Meteorology, 2: 28–31.
(in Chinese)
Ye Duzheng, 1952. The season change of the influence on Tibetan Plateau to the atmosphere circumfluence. Acta
Meteorologica Sinica, 23(1/2): 33–47. (in Chinese)
Yu Kaiju, 2004. The wavelet characteristics of near-50 year precipitation in Jingdezhen. Jiangxi Meteorology
Science, 27 (1): 23–25. (in Chinese)
Zhang Qibing, Cheng Guodong, Yao Tandong et al., 2003. A 2,326-year tree-ring record of climate variability on
the northeastern Qinghai-Tibetan Plateau. Geophysical Research Letters, 30(14): 1739–1742. (in Chinese)
Zheng Wanjun, 1983. SYLVA SINICA. Beijing: China Forestry Publishing House, 10: 289. (in Chinese)